Dendritic Arbor Developement

by Dan Peruzzi, Ph.D.

One factor that makes a neuron uniquely suited for a particular function is its morphology, including where and how the dendrites extend. The nature of the dendritic arbor affects the connectivity and electrical properties of the neuron, and arbor abnormalities are associated with neurological diseases. Many classification schemes have been based on neuronal morphology but in an article from the December 2007 issue of Neuron, “Knot/Collier and Cut Control Different Aspects of Dendrite Cytoskeleton and Synergize to Define Final Arbor Shape“, Dr. Jinushi-Nakao, Dr. Arvind, and their colleagues demonstrate a mechanism by which different classes of dendritic arbors arise.

Powerful techniques are used to take advantage of a well-defined biological system and use what is already known to carefully frame a question about how a certain type of dendritic morphology is created. In Drosophila, the cell lineage and anatomical position of dendritic arborization neurons is well known. Dendritic arborization (da) neurons develop from the external sensory organ precursor cell and are individually named. The dendritic arbors of da neurons are easily visualized; the dendrites spread out along the epidermis, practically in two dimensions, underneath the transparent larval wall of the fruit fly. There are four classes of da neurons, class I to IV in order of increasing dendritic complexity. The different morphologies are caused at least in part by different combinations of transcription factors that determine what proteins are made in the neuron by regulating which mRNAs are made from DNA. Class I cells have none of the transcription factor CUT. Class II through IV all express CUT, but the amount expressed does not correlate with the dendritic complexity; the class IV cell has less CUT than the class III cell.

Microtubules, a main component of the cytoskeleton, are present in all four cell classes, but the class III cell has filopodia that are made with another cytoskeleton component– actin. The stage was set to look for a specific component of class IV cells that makes the most complex dendritic arbors of da neurons despite
low levels of CUT. Immunohistochemistry and in-situ hybridization were used to confirm that the transcription factor, KNOT, and the mRNA for KNOT are present in only class IV neurons. Loss-of-function analysis was carried out on class IV cells with the lethal null allele for KNOT, and lethality was avoided by using mosaic analysis with a repressible cell marker in order to make fruit fly larvae that had the mutation in only some neurons. Class IV da neurons without KNOT had less complex arbors. Neurolucida was used to show significant decreases in dendritic length, number of termini (indicating less branching) and the area occupied by dendrites. There was also a change in the qualitative nature of the dendritic tree; it became more polarized. When KNOT was expressed in Class I neurons the complexity of the dendritic arbor increased.

Cytoskeleton components were also examined. First it was confirmed that CUT is needed for development of actin-rich filopodia that are present on class III cells. Labels for the whole cytoskeleton (mCD8::GFP), microtubules (FUTSCH) and a genetic construct with GFP fused to the actin binding domain were used to examine cytoskeleton changes. In wild class I neurons, all areas of the arbor had microtubules. If only KNOT was expressed in class I cells, the total dendritic arbor increased due to an increase in microtubules. If only CUT was expressed in class I cells, the total dendritic arbor increased with no increase in microtubules, but there was an increase in actin. If KNOT was expressed in a class III cell, the number of filopodia went down. Conversely, if CUT was expressed in a class IV cell and KNOT was reduced, there were many more filopodia present; the class IV cell looked more like a class III cell.

Finally, the authors searched the Gene Ontology Database looking for candidates that might be controlled by KNOT by searching for proteins associated with microtubule biogenesis and function. Spastin mRNA was up-regulated in class IV da neurons and in da neurons that were made to express KNOT, and Spastin protein was present in higher levels in class IV neurons. Interfering with Spastin mRNA in da neurons compromised dendrite outgrowth only in class IV neurons. Spastin is known to cleave microtubules and may be needed to keep microtubule growth active in class IV neurons. The human Spastin gene is mutated in many hereditary spastic paraplegia cases. The authors provide evidence that KNOT and CUT together cause greater dendritic complexity through Spastin’s affect on microtubules, and that KNOT represses CUT-mediated creation of actin-rich filopodia in class IV da neurons.

Jinushi-Nakao, S., A. Ramanathan, et al., Knot/Collier and Cut Control Different Aspects of Dendrite Cytoskeleton and Synergize to Define Final Arbor Shape. Neuron, 2007. 56: p. 963-978.

Dan Peruzzi is a staff scientist at MBF Bioscience.

First published in The Scope, fall 2008.